Design And Implementation Of Finger Vein Authentication System

With the continuous development of economy and society,more and more people pay attention to the problems in the field of information security.Biometric recognition technology has attracted the attention of industry and academia because of its high security level,convenience and other advantages.Finger vein features stand out in the field of biometric recognition with its unique body characteristics and living test advantages.In this paper,the development and design of identity authentication system based on finger vein recognition technology mainly includes vein image acquisition,vein image quality evaluation,vein image preprocessing,vein feature extraction and identity matching recognition.(1)In view of the interference of visible light and scattered light on image quality in the process of acquisition,this paper proposes to use the combination of filter and polarizer to separate the visible light.The filter is mainly responsible for separating the visible light,and the polarization device is mainly responsible for separating the scattered light.The combination of the two can effectively improve the quality of the vein image.Aiming at the phenomenon that the surface of fingers is not clean or the local brightness of fingers is not equal in the process of acquisition,this paper proposes the evaluation indexes of brightness uniformity and noise intensity,and proposes a comprehensive quality evaluation method based on MEA-BP-Ada Boost neural network in the aspect of comprehensive quality evaluation,which optimizes the initial weight and threshold of BP neural network,The decision tree is used to cascade the strong classifiers into a cascade classifier.The experimental results show that the accuracy of this method for high and low quality classification of vein images is significantly improved.(2)In order to solve the problem of unclear veins caused by noise and illumination,an improved guided filtering vein image enhancement method is proposed in this paper.The method uses sharpening to enhance the detail features of the guided image and compensate the background of the image.In addition,the filtered details are converted to logarithmic domain for nonlinear enhancement to enhance the veins and suppress the background noise,The experimental results show that the peak signal-to-noise ratio,structural similarity,information entropy and recognition rate of the enhanced vein image are improved.(3)Aiming at the problems of insufficient feature extraction and feature loss in low contrast and fuzzy vein images,this paper proposes a feature extraction method based on the fusion of local monotone pattern(LMP)in spatial domain and local phase quantization(LPQ)in frequency domain.LMP has strong robustness against the interference caused by illumination,LPQ uses the quantized Fourier image spectrum to describe the vein image in frequency domain,which has fuzzy insensitivity and effectively extracts the vein image frequency domain features.The fusion of the two can extract the vein information in both spatial domain and frequency domain,The recognition performance of identity authentication system is improved.